Fixed the links in my previous comment. In the afforementioned RUSI podcast they essentially say the same thing as you say Sept. 4th but also predict withdrawal 2 months later at best. We're 4 months in and they've made a second push (though nowhere near as hard or effective afaic see).

Point is, ~50k troops were pulled from the front line and the artillery imbalance favouring russia reduced to a quarter what it was previous to the counter-invasion. That level of redeployment alone causes disarray and fosters opportunities to take advantage of the confusion which, IMO, explain the high level of losses seen of late. If you look at how the front line moves from Toretsk and northward russian advances come to a halt (while admittedly southward/Prokrovsk is bad the whole time) for over a month before resuming at a pace nowhere near that seen during Bahkmut offensive.

We can armchair general all we want, but Kursk happened and continues to do so. Even with NK reinforcements cracks in russia's assault are showing and Kursk went a long way to making that happen which makes it a 'good' thing not even including the political and morale implications.

Side note regarding UA sourcing bias: I find you can practically modify numbers by 20% and that usually brings them to within those provided by other 'neutral' sources, so when UA says 60k russian troops moved 50k is the number I hear in my head. Other than that they are usually reliable and far better than RU sources.